How optimising for AI (artificial intelligence) differs from classic SEO

Updated 22 October 2026 · 5 min read

Classic SEO competes for a position among ten links; optimising for AI competes to get your brand into one short answer. Instead of rankings, you measure how often you are mentioned, recommended and your share of voice. The foundation is shared: an indexed site with useful content.

In classic search, people see a page of links and choose where to click. If you are in fifth place, you still have a chance of a click.

An AI assistant behaves differently. ChatGPT, Gemini and Google AI Overviews give a ready-made answer: a few sentences and a short list of options. People often don't click any further. If your brand is not on that list, as far as they are concerned you don't exist in that answer. There is no "sixth place" here.

That is why optimising for AI (also called GEO, AEO or LLMO) answers different questions: does AI mention the brand, does it recommend it, what exactly does it say about it, and which sites does it link to in the process.

Comparison: SEO vs AI optimisation

Classic SEOAI optimisation
What the user seesA list of linksA ready-made answer with a short list of options
GoalA high page rankingThe brand mentioned and recommended in the answer
Main sourcesYour siteYour site plus media, listicles, maps, reviews, competitors' sites
MetricsRankings, impressions, clicks, CTRMention rate, recommendation rate, share of voice, citations of your site
StabilityRankings change graduallyThe answer to the same query can differ from day to day
Factual accuracyYou control your snippetAI may retell facts from other sites inaccurately
ToolsSearch Console, SEO toolsServices that regularly put queries to AI assistants

Different sources

SEO works mostly with your own site: structure, content, speed and links to it. An AI answer is built differently. The model looks for information in several places and pulls it together. Google describes this as query fan-out: AI Overviews and AI Mode may run several related searches across subtopics and show a wider set of links.

In practice this means third-party sites influence the answer about your niche. In our monitoring, among AI answers with links, the brand's own site was cited in 54% of cases, and media, blogs and other sites in the same 54%. Competitors' sites appeared in 36% of answers, maps and local directories in 20%. The full table and explanation are in the article on which sources AI relies on.

The conclusion is simple: for AI, being well optimised on your own site is not enough. You need to be in the listicles, reviews and directories that AI reads in your niche.

Different metrics

In SEO you count rankings, impressions and clicks. For AI answers these metrics don't work: there is no ranking, and there isn't always a click. Instead you measure:

  • Mention rate — the share of answers in which the brand is mentioned at all.
  • Recommendation rate — the share of answers in which AI didn't just mention the brand but recommended it.
  • Share of voice — your share of all brand mentions across your queries compared with competitors.
  • Citations of your site — how often the answer links to your pages specifically.

An important point: AI answers are unstable. The same query today and tomorrow can produce different lists. So a single check proves nothing, and conclusions are drawn from the trend over a week or a month. That is exactly how AI Visibility calculates its metrics: daily checks, a trend over the period and the option to trace every number back to the real answers.

Another difference is coverage. Google doesn't show an AI Overview for every query. In our own project, for niche B2B queries, an AI Overview appeared for only 3 of 21 queries and AI Mode for 6 of 21. How Google decides when to show an AI block is covered in the article on Google AI Overviews.

What stays the same

Much of SEO works for AI too:

  • Indexing. Google says that for AI Overviews and AI Mode a page must be indexed and eligible to be shown with a snippet. There are no additional requirements.
  • Crawler access. If your site blocks search crawlers in robots.txt, AI won't link to it. For ChatGPT, the crawler is OAI-SearchBot.
  • Useful text content. Prices, addresses, terms and comparisons should be on the page as text, not only in an image.
  • Structured data. In its guide to optimising for generative AI, Google says no special markup is needed for AI but recommends continuing to use structured data as part of SEO.
  • Reputation. Reviews, media mentions and accurate directory listings matter for both channels.

In other words, SEO is the foundation, and AI optimisation adds work on third-party sources, fact control and new metrics on top of it.

What a business should do in practice

  1. Don't abandon SEO: it is the basis for both channels.
  2. Put together 10–30 queries that customers ask AI assistants, in their language and for their city.
  3. See who is mentioned in the answers and which sites they link to.
  4. Add the pages your site is missing and fix the errors AI repeats.
  5. Work on your presence in the sources AI relies on in your niche.
  6. Measure results regularly rather than with a one-off check.

How to check for your business

The quickest way is a free live check: enter your brand and a customer question, and in 20–60 seconds you'll see a real ChatGPT answer. To compare answers from ChatGPT, Gemini, Google AI Overviews and AI Mode over several days, use the free first week.

Questions and answers

Will AI optimisation replace classic SEO?

No. Google says plainly that optimising for generative search is still SEO. AI answers rely on indexed pages, so without an SEO foundation there will be no results.

Why does my site rank at the top of Google while ChatGPT never mentions me?

AI builds an answer from several sources, and often these are not your site but listicles, media, maps or reviews. If you are not there, a high ranking for your site won't help.

What is share of voice in AI answers?

It is the share of mentions of your brand among all brand mentions in the answers to your set of queries. It shows how visible you are compared with your competitors.

Do I need a dedicated specialist for AI optimisation?

Not necessarily. An SEO specialist or marketer can do it, provided they have a tool that collects AI answers regularly and time to work on sources beyond your site.

Check your business

See what ChatGPT tells your customers right now — or get a week of daily monitoring and your first action plan for free.